[PATCH] Re: netlogon_creds_cli_validate() in master4-schannel

[Garming Sam]

On 02/12/13 09:59, Andrew Bartlett wrote:
> On Fri, 2013-11-29 at 18:45 +0100, Stefan (metze) Metzmacher wrote:
>> Am 29.11.2013 06:19, schrieb Garming Sam:
>>> On 29/11/13 15:19, Andrew Bartlett wrote:
>>>> On Fri, 2013-11-29 at 14:50 +1300, Garming Sam wrote:
>>>>
>>>>>> Segmentation fault (core dumped)
>>>>>>
>>>>>>
>>>>>>
>>>>> The seg fault was because it couldn't fetch the secret. This patch
>>>>> prints an error in this case and fixes the seg fault by reinitializing
>>>>> the pointer.
>>>> Thanks Garming!
>>>>
>>>> Reviewed-by: Andrew Bartlett <abartlet at samba.org>
>>>>
>>>> Metze: can you review that and add it to you tree?
>>>>
>>>> Thanks,
>>>>
>>>> Andrew Bartlett
>>>>
>>> Andrew and I just wrote a test to cover the calls that were changed.
>> Thanks! I've integrated the patches into my tree, Partly before my changes.
> Thanks.  Can you add my review tags to the patches I reviewed last week,
> so I don't have to re-tag them?
>
> I'll also work with Garming to do some manual testing and perhaps add a
> few more automated tests.
>
> Andrew Bartlett
>

Hi there,

We were aiming to write a test to demonstrate the difference between 
master and some of your patches.

On git master:
Calling wbinfo -t, net rpc testjoin, wbinfo -t on a machine joined to a 
Windows domain gave an auth error - could not check secret. But 
attempting to do the same when joined to a Samba domain inside a test 
environment didn't appear to produce any errors.

On your branch:
Calling wbinfo -t, net rpc testjoin, wbinfo -t on a machine joined to a 
Windows domain appeared to work normally. In the Samba test environment, 
no errors appeared either.

The patch has clearly made a difference to joining a Windows domain, 
which is great. It would be good to have something like that for Samba 
to test though, any ideas?



Garming Sam